Traditional Decoction and PUAE Aqueous Extracts of Pomegranate Peels as Potential Low-Cost Anti-Tyrosinase Ingredients

The aim of the study is to evaluate the anti-tyrosinase activity of different aqueous extracts obtained from pomegranate juice processing by-products. External pomegranate peels of two certified cultivars (Akko and Wonderful), were extracted using only water as the extraction solvent. A traditional decoction and a pulsed ultrasound-assisted extraction (PUAE), both 10 min long, were performed and compared. All the aqueous extracts proved to be rich in bioactive compounds. In particular, the total phenolic content (TPC) ranged from 148 to 237 mg gallic acid equivalent (GAE)/g of dried peels (DW), the radical-scavenging ability (RSA) ranged from 307 to 472 mg ascorbic acid equivalent (AAE)/g DW, the free ellagic acid content (EA) ranged from 49 to 94 µg/mL, and the ellagitannins (ETs) ranged from 242 to 340 µg/mL. For both cultivars, PUAE extracts had higher ET content and a lower EC50, while the decoctions had slightly higher TPC, RSA, and free EA amounts. Principal component analysis (PCA) highlighted the direct correlation between the ET content and the tyrosinase enzyme inhibition (lower values of EC50). These findings suggest the potential use of both these natural extracts as low-cost lightening and/or anti-browning ingredients exploitable in several formulations (e.g., cosmetics) or extemporarily usable.


Introduction
Pomegranate (Punica granatum L.) is one of the oldest edible fruits. Mentioned in the Bible, the Koran and Egyptian texts, the first pomegranate crops date back to 3000-4000 BC. [1].
Native to the Middle East and North Africa, today pomegranate is spread globally due to its high longevity, drought and salinity resistance, and adaptability to different climatic conditions [2]. In particular, pomegranate is widely cultivated in Iran, India, China, USA, several countries of the former USSR, South Africa, and in Mediterranean countries such as Spain, Turkey, Egypt, Israel, To the best of our knowledge, no studies of the aqueous extracts of the Akko and Wonderful varieties of pomegranate as sources of anti-tyrosinase compounds have been conducted, and this is the object of our investigation.

Pomegranate Samples
Pomegranate fruits of Akko and Wonderful certified cultivars were harvested at Masseria FruttiRossi (https://lomesuperfruit.com/it/home) in Castellaneta Taranto (Italy), which includes 250 hectares of crops and represents the major Italian pomegranate producer.
Fruits were collected at full maturity (Maturity Index, MI = 14.5 ± 2.5), during September and November 2018, respectively, for Akko and Wonderful cultivars, and immediately processed to obtain the corresponding juice. Peels, manually separated during the juice processing, were dried using a traditional heating oven (Binder FED53, Goettingen, Germany) for 48 h at 40 • C. Before the extraction process, dried peels were finely ground by a Grindomi × 200 M (Retsch, Haan, Germany) for 20 s at 5000 rpm and finally sieved by a 150 µm sieve to obtain a fine and homogeneous material.

Pomegranate Peel Extraction
Peel extraction, both by traditional decoction and by 'green' extraction exploiting the ultrasound action, was performed. Only water was used as the extraction solvent under both methods. The same extraction time (10 min) and the same solvent/solid ratio (40/1, considering the dried peel weight) were used for both extraction processes. Each extraction was performed in duplicate and all the obtained extracts were analytically analyzed.

Traditional Decoction (D)
The decoction of pomegranate peels was carried out by adding the milled dried materials to 40 mL of Milli-Q water in a beaker. This suspension was heated and kept boiling for 10 min. Then, it was filtered by Buchner, using Whatman n. 1 paper, and centrifuged at relative centrifugation force (RCF or G-force) equal to 2016 for 10 min. The final extracts (AK_D and WO_D) were kept at −20 • C until further analysis.

Green Extraction: Pulsed Ultrasound-Assisted Extraction (PUAE)
Ultrasonication was applied directly by an Hielscher UP200St (Teltow, Germany) in pulsed modality (PUAE). Maximum nominal output power of 200W, frequency of 26 kHz, and a titanium probe (7 mm diameter) were employed for the sonication. The pulse duration and pulse interval refer to "on" time and "off" time of the sonicator. The pulse duration/pulse interval ratio was 4/1 (duty cycle 80%). The amplitude level was set to 50%. During the extraction process, temperature was controlled to be below 65 • C (temperature variation during the extraction period was lower than 40 • C). The obtained extracts (AK_PUAE and WO_PUAE), similarly to the decoction described in the previous paragraph, were filtered, centrifuged and stored frozen until analysis.

Characterization of the Peel Aqueous Extracts
pH was evaluated by electrochemical measurements using a pHmeter (Jenway 3510, Stone, UK). The total soluble content (TSS), expressed in • Brix, was determined at 20 • C by a digital refractometer (Hanna Instruments, Milano, Italy) ( Table 1). The total polyphenol content (TPC) and the Radical Scavenging Activity (RAS) of all the aqueous peel extracts (D and PUAE) were determined spectrophotometrically. The total anthocyanin content (TAC) and the EA and ET amounts were quantified by chromatography. TPC was determined by an UV-Vis Agilent 8453 spectrometer (Waldbronn, Germany) following the Folin-Ciocalteu method previously reported by Singleton et al. [28]. TPC was determined in the peel aqueous extracts as milligrams of gallic acid equivalent in 100 mL of extract (mg GAE/100 mL) using an appropriate gallic acid calibration curve (R 2 = 0.9887) as the reference standard. Then, the results were transformed in mg GAE/g of pomegranate peel dry weight (DW). Each sample was analyzed in duplicate and the results were reported as mean values ± standard error (SE).

Determination of the Radical Scavenging Activity
RSA of the achieved extracts was spectrophotometrically determined by DPPH• (1,1-diphenyl-2-picryl-hydrazyl) in vitro assay [29]. This method, widely used for food matrices, allowed the detection of the antiradical activity colorimetrically based on the discoloration of the stable colored DPPH• recorded at 515 nm. Ascorbic acid was used as the reference standard and the recorded absorbances of the extracts were compared with the absorbance of solutions of ascorbic acid (AA) at known concentration (y = 0.5824x − 0.0036; R 2 = 0.9975; range of linearity 0.1-1 mM). The antiradical activity was obtained as mg ascorbic acid equivalent in 100 mL of extract (mg AAE/100 mL) and then converted as mg AAE/g of pomegranate peel DW. Two replicated determinations were performed for each sample and data were reported as mean values ± SE.

Determination of EA and ETs
The HPLC determination of EA and ETs was realized using an Agilent 1100 LC-DAD (Agilent Technologies, Palo Alto, USA) [4]. A C-18 Kromasil 100 ® (Akzo-Nobel, Amsterdam, NL) column (250 × 4 mm, 5 µm), thermostat-controlled at 30 • C, was used for the chromatographic separations. Gradient elution of acetic acid/methanol/water (2:400:598) (solvent A) and methanol (solvent B), with a 1 mL/min flow rate was employed as follows: 100% A at 0 min, 0% A at 20 min, 100% A reset in 10 min and held for 5 min. The injection volume was 10 µL, and the detection wavelengths were set between 220 and 600 nm. Free EA in the extracts was quantified at 245 nm by external standard, using an appropriate 4-point calibration line (y = 251.58x − 204.41; R 2 > 0.999; range of linearity 2.5-200 µg/mL; LOD = 0.78 µg/mL; LOQ = 2.36 µg/mL). The total ET content was determined after the preliminary acid hydrolysis (performed in replicate) of extracts in order to hydrolyze ETs in EA [4,30]. A quantity of 2 mL of each extract, added to 2 mL HCl 2 M, was heated at 100 • C for 1 h in a boiling water bath. After the cooling time, 1 mL of NaOH 2M and 6 mL of methanol were added. The final solution was vortexed and filtered before the injection. Two replicated determinations of EA and ETs were performed for each sample analyzed and data were reported as mean values ± SE. The EA chromatogram of one pomegranate peel aqueous extract is reported in the Supplementary Materials Figure S1.

Determination of Total Anthocyanins
TAC identification and quantification was conducted using an HPLC-DAD method [31]. Analysis were carried out on a C-18 core-shell column (Agilent Poroshell 120 EC, 3 mm × 150 mm, 2.7 µm particle size) thermostat-controlled at 30 • C. Gradient elution of water acidified at pH 2.14 by formic acid (solvent A) and acetonitrile (solvent B), with a 0.4 mL/min flow rate, was employed as follows: 95% A at 0 min, 70% A at 25 min, 45% A at 35 min, and 0% A at 42 min, held for 3 min. Anthocyanins were quantified at 515 nm by external standard, using an appropriate 4-point cyanidin-3-O-glucoside calibration line (y = 71.52x -12.54; R 2 > 0.999; range of linearity 1.5-150 µg/mL; LOD = 0.39 µg/mL; LOQ = 1.17 µg/mL). Each sample was analyzed in duplicate and the results, expressed as µg cyanidin-3-O-glucoside/g DW, were reported as mean values ± SE. The anthocyanin profile of one pomegranate peel aqueous extract is reported in the Supplementary Materials Figure S2.

Tyrosinase Inhibition Property
A 25 mg/mL stock solution of each cultivar and treatment, prepared as described in the previous paragraph, was diluted in water to obtain a series of test solutions with final concentrations of 10, 50, 100, 250, 350, and 500 µg/mL. Components of the reaction mix were added to each well of 96-well plates in the following order: 70 µL of phosphate buffer, 60 µL of extract solutions (water for controls), 10 µL of mushroom tyrosinase (Sigma-Aldrich, T3824, 25 kU, 125 U/mL in phosphate buffer, pH 6.8), and 70 µL L-tyrosine (0.3 mg/mL in water). Kojic acid was used instead of pomegranate solutions as a positive control. Blank samples without enzyme were also included for all conditions. Plates were then incubated at 30 • C for 60 min, and absorbance was read at 505 nm in a microplate reader (Spectra Max 340PC). Percent inhibitory activity (I%) was calculated according to the formula: where Aex/en = absorbance of sample mixture with extract and enzyme; Aex = absorbance of sample mixture with extract and without enzyme; Aen = absorbance of sample mixture with enzyme and without extract; Abk = absorbance of sample mixture without enzyme and extract (blank).

Statistical Analysis
All the measurements performed were replicated and the results are expressed as mean value ± SE. Analysis of variance (ANOVA) was performed using the Excel Data Analysis Tool (Microsoft Corporation, Seattle, WA, USA). A p-value lower than 0.05 was statistically significant. Principal component analysis (PCA), based on the NIPALS algorithm [32], was performed by CAT (Chemometric Agile Tool) [33]. This chemometric software, based on R, was developed by the Chemistry Group of the Italian Chemical Society.

Results and Discussion
It is known that bioactive compounds are variable in pomegranate, and generally in plant material, according to genetic factors, different collection stages, pedoclimatic conditions of the sampling sites (growing wild or cultivation), application of agricultural practices, and post-harvest handling [34,35].
For this investigation, pomegranates of Akko and Wonderful cultivars were grown in the same cultivation area in Southern Italy, under the same pedoclimatic conditions, and using the same agro-techniques. These samples, coming from one of the biggest Italian producers, are representative of the commercial cultivars most widespread and marketed in Italy. Fruits were harvested in different collection periods according to the cultivar maturity but at the same maturity status. pH, TSS, and TAC of both the aqueous extracts (D and PUAE) of Akko and Wonderful cultivar are reported in Table 1.
One-way ANOVA, analyzing the mean differences of the pH values, showed significant differences (p < 0.05) between the cultivars (Akko and Wonderful), but no significant differences are highlighted between the two different extraction methods (D and PUAE) for the same cultivar (Table 2). As regards TSS, one-way ANOVA highlighted significant differences (p < 0.05) between D and PUAE extracts for both the cultivars. In particular, D extracts (AK_D and WO_D) showed higher TSS with respect to the corresponding PUAE extracts (AK_PUAE and WO_PUAE). This is likely because the higher boiling temperature of decoction compared to sonication allowed the partial hydrolysis of glucosides and the release of sugar moieties [36]. No significant difference between AK_PUAE and WO_PUAE was observed, probably because the extraction temperature was similarly maintained under control (<65 • C). According to previous studies [37], extraction temperature in the range of 60-70 • C was optimal for the UAE of EA from pomegranate peels.
For both the extraction methods the obtained extracts can be considered almost sugar-free if compared with the corresponding pomegranate juice (Akko: TSS = 15.20 ± 0.42; Wonderful: TSS = 17.01 ± 0.57), as previously studied by the authors [23].
Previous studies have shown that cultivar, maturity stage, and environmental factors could significantly influence the TAC of pomegranate peel extracts [5,38,39]. Cyanidin and pelargonidin derivatives are the most abundant anthocyanin pigments isolated from the peels. In particular, cyanidin 3-O-glucoside was considered one of the major anthocyanins detected in these by-products [38]. In all the extracts the TAC was slightly lower compared to those reported in literature for other pomegranate cultivars [40] even if the cited authors used methanol as an extraction solvent. One-way ANOVA, analyzing the mean differences of the TAC, showed significant differences (p < 0.05) between the cultivars when extracts were obtained by decoction. On the contrary, no significant differences between AK_PUAE and WO_PUAE were observed. Different results were shown for the two cultivars by comparing the two extraction methods (D and PUAE). Figure 1 reports the TPC and the RSA results of the peel aqueous extracts (D and PUAE) belonging to the two cultivars investigated. TPC and RSA values were moderately correlated (R 2 = 0.7304). The decoction demonstrated higher TPC compared to the extract obtained by sonication (PUAE) in both cultivars. In fact, one-way ANOVA, analyzing the mean differences of the TPCs, showed significant differences (p < 0.05) between the extraction methods (D and PUAE) for both cultivars. Analyzing the two different cultivars, there is a significant difference between AK_D and W_D, with a higher value for Akko. On the contrary, no significant differences in TPC were observed between the cultivars when the extraction was performed by PUAE.
As regards RSA, no significant differences were observed by ANOVA in Akko and Wonderful cultivars extracted both by decoction and PUAE, although the statistical significance was very low ( Table 2). Some previous studies, which considered aqueous extracts of pomegranate peels starting from different cultivars, showed an extraction efficiency, expressed as TPC, lower than that obtained in the present work; for example, El-Said et al. 2014 [41] reported a TPC equal to 16.343 mg GAE/g for aqueous extracts obtained from similarly oven-dried peels, which is more than 10 times less than those reported in Figure 1. Rahja et al. 2019 [24] described different 'green' extraction methods using just water as a solvent to efficiently extract the bioactive fraction of pomegranate peels. From a comparison of different aqueous extraction methods applied to pomegranate peels, such as ultrasound (UAE), infrared, pulsed electric fields, and high-voltage electrical discharges (HVED), HVED-assisted extraction, applied for 7 min., yielded a maximal concentration of phenolics equal to 46 mg GAE/g DW. After the same period, the UAE method yielded phenolic recovery equal to 14.5 mg of GAE/g DW, which is approximately 10-16 times lower than the TPCs of those reported in Figure 1. Recently, Kaderides et al. 2019 [42], exploiting the action of ultrasound and microwave to extract the phenolic compounds from pomegranate peels, described TPCs equal to 119.82 mg GAE/g dry peel and 199.4 mg GAE/g dry peel, respectively, which are more in accordance with the values described in this study (148-237 mg GAE/g DW). Figure 2 reports the free EA and ET amounts of the peel aqueous extracts (D and PUAE) belonging to the two different cultivars investigated. TPC and RSA values were moderately correlated (R 2 = 0.7304). The decoction demonstrated higher TPC compared to the extract obtained by sonication (PUAE) in both cultivars. In fact, one-way ANOVA, analyzing the mean differences of the TPCs, showed significant differences (p < 0.05) between the extraction methods (D and PUAE) for both cultivars. Analyzing the two different cultivars, there is a significant difference between AK_D and W_D, with a higher value for Akko. On the contrary, no significant differences in TPC were observed between the cultivars when the extraction was performed by PUAE.
As regards RSA, no significant differences were observed by ANOVA in Akko and Wonderful cultivars extracted both by decoction and PUAE, although the statistical significance was very low ( Table 2). Some previous studies, which considered aqueous extracts of pomegranate peels starting from different cultivars, showed an extraction efficiency, expressed as TPC, lower than that obtained in the present work; for example, El-Said et al. 2014 [41] reported a TPC equal to 16.343 mg GAE/g for aqueous extracts obtained from similarly oven-dried peels, which is more than 10 times less than those reported in Figure 1. Rahja et al. 2019 [24] described different 'green' extraction methods using just water as a solvent to efficiently extract the bioactive fraction of pomegranate peels. From a comparison of different aqueous extraction methods applied to pomegranate peels, such as ultrasound (UAE), infrared, pulsed electric fields, and high-voltage electrical discharges (HVED), HVED-assisted extraction, applied for 7 min., yielded a maximal concentration of phenolics equal to 46 mg GAE/g DW. After the same period, the UAE method yielded phenolic recovery equal to 14.5 mg of GAE/g DW, which is approximately 10-16 times lower than the TPCs of those reported in Figure 1. Recently, Kaderides et al. 2019 [42], exploiting the action of ultrasound and microwave to extract the phenolic compounds from pomegranate peels, described TPCs equal to 119.82 mg GAE/g dry peel and 199.4 mg GAE/g dry peel, respectively, which are more in accordance with the values described in this study (148-237 mg GAE/g DW). Figure 2 reports the free EA and ET amounts of the peel aqueous extracts (D and PUAE) belonging to the two different cultivars investigated. Free EA and ET contents were not correlated. In particular, for both cultivars, the decoction (AK_D and WO_D) showed higher EA and lower ET contents compared to the PUAE extracts (AK_PUAE and WO_PUAE). The PUAE extract, according to several other green strategies reported in the literature [42][43][44], confirms its superiority in terms of ET recovery, even if the simplest decoction is endowed with a higher free EA content, probably due to the highest boiling temperature, which allows the partial hydrolysis of ETs with the consequent release of the common aglycone (EA) and some sugar moieties [36]. The aim of this investigation was the comparison between the innovative PUAE and the traditional decoction. For this reason, the same experimental conditions (extraction time, liquid/solid ratio, extraction solvent) were maintained for both of the extraction strategies with the exception of the temperature, which was the only variable that changed (PUAE < 65 °C and D = 100 °C). ETs are stable in acidic conditions (pH = 2-4), which is their natural condition; nevertheless, they rapidly degrade in neutral and mildly basic media, especially at higher temperatures [36]. In acidic conditions, as in the pomegranate extracts, ETs show high stability in the range of temperature 20-60 °C but at 80 °C they gradually degrade [36].
One-way ANOVA, analyzing the mean differences of the free EA content, showed significant differences (p < 0.05) only between AK_D and WO_D. On the contrary, as regards the mean differences of the ET content, one-way ANOVA did not reveal significant differences between the studied couples as reported in Table 2. Figure 3 reports the percentage of tyrosinase inhibition of the peel aqueous extracts (Akko and Wonderful) obtained by the innovative (a) and the traditional (b) extraction techniques. Free EA and ET contents were not correlated. In particular, for both cultivars, the decoction (AK_D and WO_D) showed higher EA and lower ET contents compared to the PUAE extracts (AK_PUAE and WO_PUAE). The PUAE extract, according to several other green strategies reported in the literature [42][43][44], confirms its superiority in terms of ET recovery, even if the simplest decoction is endowed with a higher free EA content, probably due to the highest boiling temperature, which allows the partial hydrolysis of ETs with the consequent release of the common aglycone (EA) and some sugar moieties [36]. The aim of this investigation was the comparison between the innovative PUAE and the traditional decoction. For this reason, the same experimental conditions (extraction time, liquid/solid ratio, extraction solvent) were maintained for both of the extraction strategies with the exception of the temperature, which was the only variable that changed (PUAE < 65 • C and D = 100 • C). ETs are stable in acidic conditions (pH = 2-4), which is their natural condition; nevertheless, they rapidly degrade in neutral and mildly basic media, especially at higher temperatures [36]. In acidic conditions, as in the pomegranate extracts, ETs show high stability in the range of temperature 20-60 • C but at 80 • C they gradually degrade [36].
One-way ANOVA, analyzing the mean differences of the free EA content, showed significant differences (p < 0.05) only between AK_D and WO_D. On the contrary, as regards the mean differences of the ET content, one-way ANOVA did not reveal significant differences between the studied couples as reported in Table 2. Figure 3 reports the percentage of tyrosinase inhibition of the peel aqueous extracts (Akko and Wonderful) obtained by the innovative (a) and the traditional (b) extraction techniques. The PUAE extracts of the two cultivars showed inhibition values not significantly different at the different concentrations tested, except for those observed at 500 µg/mL. More relevant differences were found for traditional decoctions: in these extracts the Akko cultivar showed greater inhibitions at all concentrations except the highest one considered. From the inhibition values obtained from the various concentrations tested, it was possible to calculate the EC50. In general, the extracts obtained from the innovative extraction technique had better EC50 values (AK_PUAE = 117.73 µg/mL; WO_PUAE = 159.04 µg/mL) compared to traditional extracts (AK_D = 147.38 µg/mL; WO_D = 365.91 µg/mL).
In order to analyze the data using a multivariate approach, principal component analysis (PCA), an unsupervised pattern recognition technique, was applied to the data matrix D8,8, whose rows are the extracts and the columns are all the determined experimental variables. In detail, the eight extracts are those obtained by PUAE and decoction for both cultivars, in duplicate, and correspondingly categorized. PCA was performed on the autoscaled data matrix, since autoscaling, which consists of mean-centering followed by division of each column (variable) by its standard deviation, allows elaboration of multivariate data characterized by different scales and units. Figure 4 a-b shows the PCA results obtained from the above-mentioned data matrix D8,8. The PUAE extracts of the two cultivars showed inhibition values not significantly different at the different concentrations tested, except for those observed at 500 µg/mL. More relevant differences were found for traditional decoctions: in these extracts the Akko cultivar showed greater inhibitions at all concentrations except the highest one considered. From the inhibition values obtained from the various concentrations tested, it was possible to calculate the EC50. In general, the extracts obtained from the innovative extraction technique had better EC50 values (AK_PUAE = 117.73 µg/mL; WO_PUAE = 159.04 µg/mL) compared to traditional extracts (AK_D = 147.38 µg/mL; WO_D = 365.91 µg/mL).
In order to analyze the data using a multivariate approach, principal component analysis (PCA), an unsupervised pattern recognition technique, was applied to the data matrix D 8,8 , whose rows are the extracts and the columns are all the determined experimental variables. In detail, the eight extracts are those obtained by PUAE and decoction for both cultivars, in duplicate, and correspondingly categorized. PCA was performed on the autoscaled data matrix, since autoscaling, which consists of mean-centering followed by division of each column (variable) by its standard deviation, allows elaboration of multivariate data characterized by different scales and units. Figure 4 a-b shows the PCA results obtained from the above-mentioned data matrix D 8,8 . The first two principal components (PCs) together explain almost 90% of the total information of the whole data set since they visualize almost 90% of the total variance of the dataset. Figure 4a-b shows the biplot PC1 vs. PC2, which represents the scores and loadings plot, on the first two PCs, highlighting the different categorization of the extracts (a-the extracts are categorized by the different cultivar; and b-the data are categorized by the different extraction techniques).
PC1, the direction of maximum variance (60.9%), for both cultivars, discriminates the PUAE extracts from decoctions which have more positive scores on the first PC (Figure 4b). Decoction could be associated with higher °Brix, RSA, TPC, TAC, and EA (highest loadings on PC1). Nevertheless, as far as anti-tyrosinase activity is concerned, PC2, whose variance is about 29%, is the most important component, highlighting the indirect correlation between the ET content (highest loadings on PC2) and the EC50 (lowest loading on PC2). For both cultivars, PUAE extracts could be associated with the highest ET content (highest loading on PC2) and the highest anti-tyrosinase activity. AK_PUAE, which represents the extract with the highest content of ETs, is the one endowed with the highest anti-tyrosinase activity (lowest EC50). Since PCs are mathematically uncorrelated (orthogonal) variables, it seems that ET content, among all the determined variables, is the most important to determine the anti-tyrosinase activity, together with pH, whose highest value in the Akko cultivar could further positively influence the anti-tyrosinase activity [10,11]. The remaining variables, having high loadings on the PC1, do not seem correlated with the anti-tyrosinase activity.
Water was employed as an extraction solvent with the dual aim of achieving a greener and safer chemistry approach compared to methanol or other organic solvents, and developing a low-cost recovery strategy of these by-products. The aqueous decoction of pomegranate peels represents the simplest and cheapest strategy to rapidly obtain extracts, and has widely been used as a folk remedy in traditional medicine. In fact, due to its high concentration of tannins, which gives the peel a strong astringency, the peel decoction has been used to treat dysentery and other gastro-intestinal disorders [45]. Moreover, another advantage of decoction is that, being an extemporaneous preparation, there is no need to stabilize the extract. This is particularly relevant in the case of the extract obtained with ultrasound, since the stability over time of aqueous extracts is very limited.
Ultrasound-assisted extraction (UAE) is a green and energy-saving technology which is being increasingly employed in the extraction of natural products as a simple and efficient alternative to conventional extraction techniques. Substantial shortening of extraction time, reduction of reagent The first two principal components (PCs) together explain almost 90% of the total information of the whole data set since they visualize almost 90% of the total variance of the dataset. Figure 4a,b shows the biplot PC1 vs. PC2, which represents the scores and loadings plot, on the first two PCs, highlighting the different categorization of the extracts (a-the extracts are categorized by the different cultivar; and b-the data are categorized by the different extraction techniques).
PC1, the direction of maximum variance (60.9%), for both cultivars, discriminates the PUAE extracts from decoctions which have more positive scores on the first PC (Figure 4b). Decoction could be associated with higher • Brix, RSA, TPC, TAC, and EA (highest loadings on PC1). Nevertheless, as far as anti-tyrosinase activity is concerned, PC2, whose variance is about 29%, is the most important component, highlighting the indirect correlation between the ET content (highest loadings on PC2) and the EC50 (lowest loading on PC2). For both cultivars, PUAE extracts could be associated with the highest ET content (highest loading on PC2) and the highest anti-tyrosinase activity. AK_PUAE, which represents the extract with the highest content of ETs, is the one endowed with the highest anti-tyrosinase activity (lowest EC50). Since PCs are mathematically uncorrelated (orthogonal) variables, it seems that ET content, among all the determined variables, is the most important to determine the anti-tyrosinase activity, together with pH, whose highest value in the Akko cultivar could further positively influence the anti-tyrosinase activity [10,11]. The remaining variables, having high loadings on the PC1, do not seem correlated with the anti-tyrosinase activity.
Water was employed as an extraction solvent with the dual aim of achieving a greener and safer chemistry approach compared to methanol or other organic solvents, and developing a low-cost recovery strategy of these by-products. The aqueous decoction of pomegranate peels represents the simplest and cheapest strategy to rapidly obtain extracts, and has widely been used as a folk remedy in traditional medicine. In fact, due to its high concentration of tannins, which gives the peel a strong astringency, the peel decoction has been used to treat dysentery and other gastro-intestinal disorders [45]. Moreover, another advantage of decoction is that, being an extemporaneous preparation, there is no need to stabilize the extract. This is particularly relevant in the case of the extract obtained with ultrasound, since the stability over time of aqueous extracts is very limited.
Ultrasound-assisted extraction (UAE) is a green and energy-saving technology which is being increasingly employed in the extraction of natural products as a simple and efficient alternative to conventional extraction techniques. Substantial shortening of extraction time, reduction of reagent consumption, increase of extraction yield, and faster kinetics are the most important advantages of UAE . The sonication used in pulsed mode (PUAE) drastically reduces the operating temperature, allowing the extraction of thermolabile compounds. The superiority of PUAE to produce antioxidants from pomegranate peels has been previously demonstrated [46]. Longer times of extraction have not been tested in the present paper in order to minimize the potential degradation of the aqueous extracts. In fact, during the sonication process of vegetal material in water, a small amount of active oxidizing species (i.e., HO· and hydrogen atoms) are generated, and this process is time-dependent [47]. Although these aqueous extracts may require a further formulation step for preservation over time [48], we did not tackle this point, since our priority was to emphasize whether or not they deserve further formulative investigations.

Conclusions
The results of this research highlighted that water, as an extraction solvent, was able to extract, in a limited extraction time (10 min), significant amounts of bioactive compounds, particularly ETs, from pomegranate peels of the commercial cultivars most widespread in Italy: Akko and Wonderful. This is of great importance considering the need for 'greener' and safer extractions of natural compounds.
Compared to PUAE, traditional decoction has the advantages of being easily and extemporaneously obtainable and leads to extracts endowed with higher values of EA, TPC, RSA, and TAC for both the cultivars investigated. Nevertheless, PUAE has the advantage of yielding extracts endowed with a higher content of ETs, whose content is directly correlated with enzymatic activity.
Based on the results obtained, both of the 'green' extraction strategies proposed proved to be promising methods to produce low-cost anti-tyrosinase ingredients. EA is already approved as a lightening/whitening agent in cosmetic formulations; however, considering these results, the cheaper natural extract rich in ETs could represent a potential new cosmetic ingredient. Stabilization to preserve the polyphenolic compounds over time and formulation of a hydrophilic gel containing the pomegranate peel extract, as well as the tests necessary to evaluate their in vivo lightening properties, are currently under study.
These findings represent an important aim for pomegranate juice manufacturers, since peels are the major form of processing waste, and their disposal is expensive and not eco-friendly.
Funding: This work was partially funded by Duferco Engineering SPA.